Abstract. Multiscale error diffusion (MED) is superior to conventional error diffusion algorithms as it

-An adaptive palette reordering method is proposed in this paper to reshape the statistical properties of the color index map of a color-indexed image with a dynamic palette. Unlike other reordering methods, the proposed method extracts information from both the color index map and the palette to achieve the objective. The compression performance of JPEG-LS can be significantly improved when the proposed method is used.

Multiscale error diffusion (MED) digital halftoning technique outperforms classical conventional error diffusion techniques as it can produce a directionalhysteresis-free bi-level image. However, extremely large computation effort is required for its implementation. In this paper, a fast MED-based digital halftoning technique is proposed to produce a halftone image without directional hysteresis at a significantly reduced computational cost. The amount of reduction is monotonic increasing with the image size. For an image of size 512x512, the proposed algorithm can save 40% of arithmetic operations as compared with MED. Moreover, since it supports parallel processing, processing time can further be squeezed.

Restoration of color quantized images is rarely addressed in the literature especially when the images are color quantized with halftoning. Conventional color image restoration algorithms are not suitable for handling this problem. A new restoration algorithm based on genetic algorithm is presented in this paper. This algorithm makes use of the available color palette and the mechanism of a halftoning process to derive useful a priori information for restoration. Simulation results showed that it was able to provide a very promising restoration result.

-Multiscale error diffusion (MED) is superior to conventional error diffusion algorithms as it can eliminate directional hysteresis completely. However, the complexity of this frame-oriented process is much higher and makes it not suitable for real-time applications. In this paper, a fast MED algorithm is proposed. The complexity of this algorithm is remarkably reduced as compared with conventional MED algorithms. It also supports parallel processing.